|
To: Fellow Researchers
From: Lynne C. Jones
I am often asked how to analyze the results of an experiment. It has become apparent that there is frequently a lack of exposure to experimental design and statistical analysis in our education. I cannot recommend strongly enough that it would behoove each and every one of us to take advantage of the statistics courses offered by this university.
I do not believe that every one of us should become "statistical experts." Formal programs in statistical analysis take several years. However, an appreciation for statistical methods would enable us to perform many of the basic tests used – such as the Student's t test, linear regression, etc. But more importantly, familiarity with statistical analysis would allow us to communicate better with the professional statistician. I cannot count how many times that someone has come to me to analyze their results when they didn't even know what question they were trying to answer.
The statistician expects you to be an authority on your results:
- What is your hypothesis? Why?
- How did you measure the outcome?
- How many repeated measurements did you make?
A lot of these questions really should be answered prior to initiating your study. Contact with a statistician, at that time may make life a lot easier at the conclusion of your study. Therefore, I strongly recommend asking a statistician, or someone with a strong background in statistics, to help you in establishing your experimental design.
Contained in this section of the website are a number of reprints of manuscripts which have focused on statistical methods in orthopaedics. As most are not written by statisticians, these articles do not displace the need for a good statistical textbook for theory and mathematics. However, these articles are written in laymen's language and address problems frequently observed in the course of conducting and publishing orthopaedic research.
Finally, within our institution there are several individuals who would be happy to assist you in the design and analysis of your experiments. Do not hesitate to use this valuable resource. I would be happy to review your project with you at any time. If I can help, I will. If I can't help, I'll direct you to someone who can.
Good Luck.
Lynne C. Jones, Ph.D.
Associate Professor
Department of Orthopaedic Surgery
Johns Hopkins University
|
